AnnabanAI

company

AI & ML interests

Foundation models (LLMs, VLMs, multimodal) Efficient training and inference (quantization, distillation, sparsity) Retrieval-augmented generation (RAG) Agentic systems and orchestration Continual and lifelong learning Systems & Infrastructure Deterministic and reproducible ML Model governance, auditability, and provenance MLOps / LLMOps (CI, evaluation, monitoring) On-device and edge inference Secure enclaves and confidential computing Safety, Trust, and Governance Alignment and controllability Interpretability and explainability Evaluation frameworks and benchmarks Tamper-evident logging and evidence systems Privacy-preserving ML (federated learning, differential privacy) Applied Domains Developer tools and copilots Scientific ML (biology, materials, climate) Data analytics and autonomous data agents Cybersecurity and fraud detection Robotics and embodied AI Optimization & Methods Reinforcement learning (offline, RLHF/RLAIF) Causal inference Graph ML Time-series and forecasting Bayesian methods Emerging Directions Multi-agent coordination Program synthesis and tool-using models Neuro-symbolic approaches Evaluation under distribution shift Energy-efficient AI